AI hotel reviews summaries misrepresent safety, Which? finds

What's happened

Which? has tested Tripadvisor’s AI hotel review summaries and found they downplay serious safety and hygiene complaints. Across Cape Verde, the Dominican Republic and Turkey, guests report food-poisoning, lack of running water and sexual harassment, while AI overviews label some hotels as “spotless” and “friendly.” Tripadvisor says summaries are snapshots, not replacements for full reviews, and cites safeguards.

What's behind the headline?

Brief

The investigation reveals a tension between AI-generated summaries and user experiences shared in reviews.
It challenges the idea that AI overviews provide a full picture of a hotel’s safety/high hygiene.

What to watch

The reliability of AI summaries as a first filter for travellers.
How Tripadvisor’s safeguards address safety warnings and what happens when reviews detail extreme harm.

Implications

Readers should cross-check AI summaries with individual reviews, especially in cases of safety concerns.
The piece raises questions about the responsibility of platforms to present accurate, non-misleading previews.

How we got here

Which?’s investigation examines how AI-generated summaries on Tripadvisor pages pull from large volumes of user reviews. It highlights inconsistencies between highlighted summaries and on-site guest reports, and notes Tripadvisor’s claim that AI summaries are not a substitute for the full set of reviews.

Our analysis

Independent and The Guardian report on Which? findings, with quotations from Which? Travel and Tripadvisor spokespeople. The Guardian cites Rory Boland and Duncan Brumby, while The Independent compiles guest accounts of hygiene issues and safety incidents.

Go deeper

Will readers trust AI summaries after this investigation?
How might Tripadvisor refine its AI safeguards to prevent misrepresentation?
What should travellers do to verify a hotel’s safety beyond summaries?